212 C. Davison — On Earthq^take- Sounds. 



caused by the slip of a fault running approximately north-east and 

 south-west, and hading to the north-west, the slip extending 

 probably over a horizontal distance of several miles, and being 

 greatest towards the south-west. Now, " the 25 places at which the 

 sound preceded, or preceded and accompanied, the shock, though not 

 confined to any one part of the disturbed area, are mostly situated 

 in the district north-east of the epicentruni," while five of the six 

 places at which the sound followed, or accompanied and followed, 

 the shock are "either south-west of the epicentrum, or close to it 

 on the north-west side, i.e. just where the intensity is greatest." 



2. The Extent of the Sound-area is independent of that of the 

 Disturbed Area. — Humboldt, in his "Cosmos," remarks that "the 

 intensity of the hollow noise which generally accompanies an earth- 

 quake does not increase in the same degree as the force of the 

 oscillations;"' and it has also been observed that, in very violent 

 earthquakes, the sounds are confined to a comparatively small area 

 in the neighbourhood of the epicentrum. The Neapolitan earthquake 

 of 1857, for instance, disturbed the whole of the Italian peninsula 

 south of lat. 42°, being felt nearly as far as Rome, while the sounds 

 were only heard within an area containing 3300 square miles imme- 

 diately surrounding the epicentrum. 



The slighter shocks of this country also afford good examples. In 

 the East Cornwall earthquake of October 7, 1889, and the Inverness 

 earthquake of November 15, 1890, the sounds were heard at nearly 

 all the places where the shocks were felt; though it is of course 

 possible that in these cases the coincidence of the sound-area and 

 the disturbed area may have been apparent rather than real. The 

 Edinburgh earthquake of January 18, 1889, disturbed an elliptical 

 area, about 30 miles from north to south, and 26^ miles from east 

 to west ; the length of the sound-area was about 25 miles from north 

 to south, its breadth could not be exactly determined. The dis- 

 turbed area of the Lancashire earthquake of February 10, 1889, was 

 approximately circular, being 56 miles from north to south and 54 

 miles from east to west ; the sound-area was very nearly circular 

 and 29 miles in diameter. 



These examples are sufficient to show that the extent of the sound- 

 area bears no constant relation to that of the disturbed area. As a 

 general rule, we may say that, the more intense the earthquake, the 

 less is the ratio of the extent of the sound-area to that of the 

 disturbed area ; but this is by no means always true. 



The limiting case, in which sounds are heard without any 

 perceptible shock, is one of which records are frequent. One of 

 the most remai'kable is that described by Humboldt in his " Cosmos," 

 ■where he refers to the subterranean thunderings [bramidos y truenos 

 anbterraneos) of Guanaxuato on the Mexican plateau. "The noise," 

 he says, "began about midnight, on the 9th of January, 1784, 



and continued for a month From the 13th to the IGth of 



January, it seemed to the inhabitants as if heavy clouds lay beneath 

 their feet from which issued alternate slow rolling sounds and short 



1 Vol. i. p. 203. 



